literature-review

Start by framing an answerable question, then define inclusion and exclusion rules before searching. Search in expanding rings with a query log, triage sources in two passes, grade each on rigor and relevance, and synthesize findings in a matrix organized by claims rather than by source.

Stop at saturation: when new sources keep citing what you have already read and add no new claims to your synthesis matrix. Define a stop-question upfront, timebox the effort, and ship with a freshness date and re-check trigger.

A synthesis matrix is a table with sources as rows and your question's aspects (method, context, finding, effect size, limitations, grade) as columns. It mechanically surfaces agreements, conflicts explained by moderating variables, and gaps that prose summaries hide.

Use technology evaluation when deciding whether to adopt a specific tool or library. A literature review maps a knowledge landscape and what is known about a topic, rather than producing an adopt-or-reject verdict on one option.

You are likely searching only for confirmation. Search the negation explicitly with queries like "X limitations" or "X replication failure" — every live topic has disagreement, and finding none means the search was incomplete.